Greetings everybody this is my first post.
Ive recently taken an interest in roulette & i hope i can contribute something of value on this forum but i feel that i may be resurrecting discarded methods of playing.
can someone please clarify the testing procedure ?
Heres a scenario
Im at a Ficticious casino & their are 10 people at a busy table, the croupier spins the wheel once every few mins.
I go away & return 1hr later ,glance at the marqee & i witnesss an unbroken sequence of 0-36 in 37 spins
several weeks week later someone rediscovers a method & tests it using past spins from that event.
The tester has time travelled & is now the 11th person betting in that event & so that 37 spin 0-36 sequence would most likely not have existed.
why do people test using past spins from live play where additional human interaction would have altered the results?
Ive used a small sample size of 37 spins to illustrate but it could easily be a million spins broken into smaller sessions
The tester was not part of that million spin event but uses the results to examine, replicate,discredit a method of play.
shouldn't multiple tests from all sources . Rng ,live,airball etc be conducted individually & the results compared to examine a method over say 1000 spins?
Ultimately the real question is what exactly are you testing. If you are testing a system or bet then you would have to test it over a large number of trials to see if there is anything there. Definately more than 1000. If you are trying to beat RNG, then test on rng. If you are trying to beat live then test on live. When it comes to testing we are talking about tens of thousands of trials and often times hundred thousands. And yes there is a difference as to how your results will vary from simple testing to times when you are actually betting. It is a good idea to seperately track the times you actually place bets. Tracking trials when you say "I should be betting" vs. times you actually place bets are two totally different events.
Thank you davey j. . . . . . . . . . . . . . . . . . . . . . . . but Im still confused .
The tester using past spin results as trial data has time travelled & would have to buy in at that table , select a chip amount & colour then bet or wait for the ideal moment . wouldn't the past spin data be void ? It may not make a difference depending on bet selection , large sample size of 100000000000 spins & method of play but isnt it still an inaccurate method or am i reading too much into this ?
Hello Element. Welcome to the forum family.
I get your point. Past results are oficially accepted not to bear any connection with future spins.
Let's say testing is empirically based on the assumption of numerical cycles and the "history repeats itself" type of observations. The "time traveller" as you put, is trying to use past spins for finding patterns/traits which may be similar to those in the future via these framework of observation.
I.e. Even if the numbers are not exactly the same, the underlying "events" may be.
For instance: numbers "1, 9, 12" equals the event "three times on a dozen", at a latter date numbers "32, 36, 28" equals the same event: "three times on a dozen". These is the type of "underlying events" fallacy-based players try to correlate to the cyclical nature of the game for future exploitation.
What is true? What really happens? I recently mentioned both happened: disconnected links for those who only see disconnection among spins, and similar traits spotted at the cyclical samples for those who care to see them.
Regards.
Victor
Hi Element,
Past spins or not, as long as they're RANDOM they can be used for testing, provided the spin sample size used is
large enough (only spins where there was a bet placed count).
Quote from: Element on September 22, 2009, 02:37:34 PM
Ive used a small sample size of 37 spins to illustrate but it could easily be a million spins broken into smaller sessions
The tester was not part of that million spin event but uses the results to examine, replicate,discredit a method of play.
shouldn't multiple tests from all sources . Rng ,live,airball etc be conducted individually & the results compared to examine a method over say 1000 spins?
Hmm, again, if the trials are random, it doesn't matter where they come from.
I think what you're saying is: The method creator could reverse-engineer the testing spins to suit his method, then present the results showing that his methods wins (with the truth being that, while it has been fit to these past trials, it can't overcome future/unknown spins).
If that's what you're saying then I would say that fitting a method to a set of spins is possible with a short-term set, but impossible with one that is considered long-term (global randomness).
In other words, you can take a session of 100 spins where there is 60 reds and 40 blacks and create a method consisting of flat-betting on Red, backing it up with your test results (that 100 spin session, which of course doesn't mean anything).
But if you take a set of a million random spins (regardless of the source and time), you can't fit any method to that in such a way that would yield a significant profit (in relation to the huge sample size).